Elucidation of a low spin cobalt(II) system in a distorted tetrahedral geometry

Article abstract of DOI:10.1021/ja026433e

We have prepared a series of divalent cobalt(II) complexes supported by the [PhBP₃] ligand ([PhBP₃] = [PhB(CH₂PPh₂)₃]^-) to probe certain structural and electronic phenomena that arise from

Full text of DOI:10.1021/ja026433e

Published on Web 11/28/2002
Elucidation of a Low Spin Cobalt(II) System in a Distorted
Tetrahedral Geometry
David M. Jenkins, Angel J. Di Bilio, Matthew J. Allen, Theodore A. Betley, and
Jonas C. Peters*
Contribution from the DiVision of Chemistry and Chemical Engineering, Arnold and Mabel
Beckman Laboratories of Chemical Synthesis, California Institute of Technology,
Pasadena, California 91125
Received April 5, 2002. Revised Manuscript Received September 27, 2002
Abstract: We have prepared a series of divalent cobalt(II) complexes supported by the [PhBP3] ligand
([PhBP3] ) [PhB(CH2PPh2)3]-) to probe certain structural and electronic phenomena that arise from this
strong field, anionic tris(phosphine) donor ligand. The solid-state structure of the complex [PhBP3]CoI (1),
accompanied by SQUID, EPR, and optical data, indicates that it is a pseudotetrahedral cobalt(II) species
with a doublet ground statesthe first of its type. To our knowledge, all previous examples of 4-coordinate
cobalt(II) complexes with doublet ground states have adopted square planar structure types. Complex 1
provided a useful precursor to the corresponding bromide and chloride complexes, {[PhBP3]Co(µ-Br)}2,
(2), and {[PhBP3]Co(µ-Cl)}2, (3). These complexes were similarly characterized and shown to be dimeric
in the solid-state. In solution, however, the monomeric low spin form of 2 and 3 dominates at 25 °C. There
is spectroscopic evidence for a temperature-dependent monomer/dimer equilibrium in solution for complex
3. Furthermore, the dimers 2 and 3 did not display appreciable antiferromagnetic coupling that is typical of
halide and oxo-bridged copper(II) and cobalt(II) dimers. Rather, the EPR and SQUID data for solid samples
of 2 and 3 suggest that they have triplet ground states. Complexes 1, 2, and 3 are extremely oxygen
sensitive. Thus, stoichiometric oxidation of 1 by dioxygen produced the 4-coordinate, high spin complex
[PhB(CH2P(O)Ph2)2(CH2PPh2)]CoI, (4), in which the [PhBP3] ligand had undergone a 4-electron oxidation.
Reaction of 1 with TlOAr (Ar ) 2,6-Me2Ph) afforded an example of a 4-coordinate, high spin complex,
[PhBP3]Co(O-2,6-Me2Ph) (5), with an intact [PhBP3] ligand. The latter two complexes were spectroscopically
and structurally characterized for comparison to complexes 1, 2, and 3. Our data for these complexes
collectively suggest that the [PhBP3] ligand provides an unusually strong ligand-field to these divalent cobalt
complexes that is chemically distinct from typical tris(phosphine) donor ligand sets, and distinct from tridentate
borato ligands that have been previously studied. Coupling this strong ligand-field with a pronounced axial
distortion away from tetrahedral symmetry, a geometric consequence that is enforced by the [PhBP3] ligand,
provides access to monomeric [PhBP3]CoX complexes with doublet rather than quartet ground states.
I. Introduction
desire to make use of these elementary principles to rationally
design catalysts selective for specific transformations motivates
ongoing interest in this area.6,7 In this paper, we suggest that
coupling (i) an axial distortion in a pseudo-tetrahedral cobalt(II)
complex of approximate 3-fold symmetry with (ii) a strong
ligand-field donor strength can provide access to a doublet rather
than a quartet electronic ground state. This result is of interest
because, to the best of our knowledge, all previously character-
ized 4-coordinate cobalt(II) systems that are low-spin have
adopted approximate square planar structure types.8,9 Owing to
its historically well-behaved and rich spectroscopy, there has
Metal centers that reside in unusual coordination geometries
sometimes display unique physical properties that correlate to
novel modes of chemical reactivity. Metalloenzymes can exploit
subtle structure/function relationships to achieve specific cata-
lytic transformations by intimately tuning the local stereochem-
istry and ligand-field in a protein active site.1 An appreciation
of specific ligand-to-metal interactions in protein active sites is
therefore highly dependent on our basic understanding of
elementary stereochemical and ligand-field relationships in
coordination chemistry.2-5 Moreover, our community’s general
(6) Schultz, B. E.; Gheller, S. F.; Muetterties, M. C.; Scott, M. J.; Holm, R.
H. Acc. Chem. Res. 1986, 19, 363-370.
(1) Lippard, S. J.; Berg, J. M. Principles of Bioinorganic Chemistry; University
Science Books: Mill Valley, CA, 1994; Chapter 12.
(7) Sellmann, D.; Sutter, J. Acc. Chem. Res. 1997, 30, 460-469.
(8) (a) Everett, G. W.; Holm, R. H. J. Am. Chem. Soc. 1965, 87, 5266-5267.
(b) Cotton, F. A.; Soderberg, R. H. J. Am. Chem. Soc. 1962, 84, 872-873.
Also see the general references 21, 22, and 38.
(2) (a) Halpern, J. Science 1985, 227, 869-875. (b) Geno, M. K.; Halpern, J.
J. Am. Chem. Soc. 1987, 109, 1238-1240.
(3) Perutz, M. F.; Fermi, G.; Luisi, B.; Shaanan, B.; Liddington, R. C. Acc.
Chem. Res. 1987, 20, 309-321.
(9) Stereochemical tuning of 4-coordinate structure types of cobalt(II), and
the ground spin-state consequences of such tuning, has been recently
described: Jaynes, B. S.; Doerrer, L. H.; Liu, S.; Lippard, S. J. Inorg. Chem.
1995, 34, 5735-5744.
(4) Raphael, A. L.; Gray, H. B. J. Am. Chem. Soc. 1991, 113, 1038-1040.
(5) Garret, T. P. J.; Glingeleffer, D. J.; Guss, J. M.; Rogers, S. J.; Freeman, H.
C. J. Biol. Chem. 1984, 259, 2822-2825.
9
15336
J. AM. CHEM. SOC. 2002, 124, 15336-15350
10.1021/ja026433e CCC: $22.00 © 2002 American Chemical Society

Low Spin Cobalt(II) System
A R T I C L E S
Scheme 1
been longstanding interest in using cobalt(II) substitution to
probe local stereochemical environments in the active sites of
native enzymes. Perhaps the most familiar example of this use
is that of zinc carbonic anhydrase.10 Establishing that pseudo-
tetrahedral cobalt(II) can exhibit spectroscopic features con-
sistent with a doublet rather than a quartet ground state is
particularly interesting in this latter context.
The series of paramagnetic cobalt complexes described below
is well suited to probing the ligand-field donor strength that is
imposed by the tris(phosphino)borate ligand [PhB(CH2PPh2)3]-,
abbreviated throughout as [PhBP3].11,12 The recently introduced
[PhBP3] ligand13 is a structurally similar but anionic relative to
the well-known tris(phosphine) ligand CH3C(CH2PPh2)3 (ab-
breviated as triphos herein)14-16 that was originally reported by
Hewertson and Watson in 1964.17 Our continued focus on
developing neutral complexes that feature a partially insulated
borate counteranion, fixed at a short but remote distance from
a transition metal that is coordinated by neutral amine or
phosphine donor arms, is in part due to the promise these formal
zwitterions hold for catalytic applications.18,19 Studying the
elementary and catalytic reaction processes of zwitterions of
this type, and comparing their reactivity to isostructural but more
conventional cationic relatives, is paramount to defining the role
zwitterionic systems might offer to homogeneous catalysis. It
is also of interest to examine the impact that the fastened borate
unit can have upon the intimate electronic structure of the
coordinated metal center. This study describes our initial efforts
to address this latter issue. We provide evidence to show that,
when coordinated to a cobalt(II) ion, the [PhBP3] anion provides
access to a unique low spin cobalt(II) complex, [PhBP3]CoI (1),
whose stereochemical structure is best regarded as distorted
tetrahedral.20 Given the intense spectroscopic and magnetic
scrutiny divalent cobalt has received during the past several
decades,21,22 elucidation of this low spin system is particularly
interesting. Complex 1 and its chloride and bromide relatives
are structurally related to the well-known but cationic, triphos-
supported cobalt(II) systems popularized by Sacconi and more
recently by Huttner.23,24 The data presented in this paper affords
a first comparative glance at the dramatic electronic conse-
quences that arise when the borate counteranion is embedded
within the phosphine donor ligand framework.
(10) (a) Solomon, E. I.; Rawlings, J.; McMillin, J. R.; Stephens, P. J.; Gray, H.
B. J. Am. Chem. Soc. 1976, 98, 8046-8048. (b) Bertini, I.; Luchinat, C.
Acc. Chem. Res. 1983, 16, 272-279. (c) Bertini, I.; Lanini, G.; Luchinat,
C. J. Am. Chem. Soc. 1983, 105, 5116-5118. (d) Khalifah, R. G.; Rogers,
J. I.; Harmon, P.; Morely, P. J.; Carroll, S. B. Biochemistry 1984, 23, 3129-
3136. (e) Briganti, F.; Pierattelli, R.; Scozzafava, A.; Supuran, C. T. Eur.
J. Med. Chem. 1996, 31, 1001-1010.
(11) Cobalt complexes supported by tris(pyrazolyl)borate ligands have been
studied extensively. For recent examples, see: (a) Detrich, J.; Konee`ny-
Vetter, W. M.; Doren, D.; Rheingold, A. L.; Theopold, K. H. J. Am. Chem.
Soc. 1996, 118, 1703-1712. (b) Jewson, J. D.; Liable-Sands, L. M.; Yap,
G. P. A.; Rheingold, A. L.; Theopold, K. H. Organometallics 1999, 18,
300-305. (c) Thyagarajan, S.; Incarvito, C. D.; Rheingold, A. L.; Theopold,
K. H. Chem. Commun. 2001, 2198-2199.
(12) Cobalt complexes supported by tris(thioether)borate ligands have been
studied recently. See: Schebler, P.; Mondimutsira, B.; Riordan, C. G.;
Liable-Sands, L. M.; Incarvito, C. D.; Rheingold, A. L. J. Am. Chem. Soc.
2001, 123, 331-332.
(13) (a) Peters, J. C.; Feldman, J. D.; Tilley, T. D. J. Am. Chem. Soc. 1999,
121, 9871-9872. (b) Barney, A. A.; Heyduk, A. F.; Nocera, D. G. Chem.
Commun. 1999, 2379-2380.
(14) The abbreviation triphos is used herein to designate the tripodal, tridentate
ligand RC(CH2PPh2)3 (R ) alkyl or aryl). This abbreviation is often used
to delineate tridentate phosphines of the type PhP(CH2PPh2)2. In this
manuscript, we restrict our use of the term triphos to represent the former
ligand class.
(15) (a) Benelli, C.; DiVaira, M.; Noccioli, G.; Sacconi, L. Inorg. Chem. 1977,
16, 182-187. (b) Ghilardi, C.; Midollina, S.; Orlandini, A.; Sacconi, L.
Inorg. Chem. 1980, 19, 301-306. (c) Dapparto, P.; Midollini, S.; Orlandini,
A.; Sacconi, L. Inorg. Chem. 1976, 15, 2768-2744.
The synthetic, structural, and spectroscopic data for the title
complex 1 are presented below to corroborate its low spin
assignment. Two related low spin cobalt(II) derivatives, {[PhBP3]-
CoIIX} (X ) Br, Cl), are also described. These latter two
complexes feature an added complexity in that they are dimeric
in the solid-state, but predominantly monomeric in solution. The
collection of comparative solid-state and solution EPR spectra
for 1, 2, and 3 is presented. For comparison, structural and
spectroscopic data for two high spin derivatives that were
obtained directly from 1 are also described.
II. Results
IIa. Synthesis and Solid-State Structures of {[PhBP3]-
CoIIX} (X ) I, Br, Cl). Scheme 1 presents the synthesis of the
three halide derivatives {[PhBP3]CoIIX} (X ) I, Br, Cl). The
high yield synthesis of the iodide complex [PhBP3]CoI, (1),
derived from the thallium reagent [PhBP3]Tl, has been previ-
ously described.18c Although the bromide and chloride deriva-
tives could be similarly prepared, they were best derived
(16) For recent papers dealing with (triphos)Co2+ chemistry, see: (a) Heinze,
K.; Huttner, G.; Walter, O. Eur. J. Inorg. Chem. 1999, 593-600. (b) Heinze,
K.; Huttner, G.; Schober, P. Eur. J. Inorg. Chem. 1998, 183-189. (c)
Heinze, K.; Huttner, G.; Zsolnai, L. Z. Naturforsch. (B) 1999, 54, 1147-
1154. (d) Winterhalter, U.; Zsolnai, L.; Kirscher, P.; Heinz, K.; Huttner,
G. Eur. J. Inorg. Chem. 2001, 89-103. (e) Rupp, R.; Huttner, G.; Kirscher,
P.; Soltek, R.; Bu¨chner, M. Eur. J. Inorg. Chem. 2000, 1745-1757.
(17) Hewertson, W.; Watson, H. R. J. Chem. Soc. 1962, 1490-1494.
(18) (a) Lu, C. C.; Peters, J. C. J. Am. Chem. Soc. 2002, 124, 5273-5273. (b)
Thomas, J. C.; Peters, J. C. J. Am. Chem. Soc. 2001, 123, 5100-5101. (c)
Shapiro, I. R.; Jenkins, D. M.; Thomas, J. C.; Day, M. W.; Peters, J. C.
Chem. Commun. 2001, 2152-2153. (d) Jenkins, D. M.; Betley, T. A.;
Peters, J. C. J. Am. Chem. Soc. 2002, 124, 11238-11239.
(20) This finding was briefly described in a previous communication. See ref
18c.
(21) For a lucid discussion of cobalt(II) EPR spectroscopy, see: Bencini, A.;
Gatteschi, D. Trans. Metal Chem. 1982, 8, 97-119.
(22) A thorough discussion of the electronic structure of cobalt(II) systems in
4-, 5-, and 6-coordinate geometries has been provided by Lever: Lever,
A. B. P. Inorganic Electronic Spectroscopy, 2nd ed.; Elsevier: New York,
1984; pp 480-505.
(23) (a) Sacconi, L.; Midollini, S. J. Chem. Soc., Dalton. Trans. 1972, 1213-
1216. (b) Dapporto, P.; Midollini, S.; Sacconi, L. Inorg. Chem. 1975, 14,
1643-1650. (c) Mealli, C.; Midollini, S.; Sacconi, L. Inorg. Chem. 1975,
14, 2513-2521. (d) Ghilardi, C. A.; Midollini, S.; Sacconi, L. J. Organomet.
Chem. 1980, 186, 279-287.
(19) We have recently prepared the bis(amino)borate ligand [Ph2B(CH2NMe2)2]-.
Betley, T. A.; Peters, J. C. Inorg. Chem. 2002, in press.
(24) See Heinze, K.; Huttner, G.; Zsolnai, L.; Schober, P. Inorg. Chem. 1997,
36, 5457-5469.
9
J. AM. CHEM. SOC. VOL. 124, NO. 51, 2002 15337

A R T I C L E S
Jenkins et al.
Figure 1. Displacement ellipsoids are represented at 50%. Structures on the right are shown without aryl carbon atoms for clarity and the dimers are rotated
such that the bridging halide ligands are eclipsed. (A) [PhBP3]CoI, (1); Selected interatomic distances (Å) and angles (deg) for each independent molecule
in crystal structure: Co(1)-P(1), 2.200(2); Co(1)-P(2), 2.206(2); Co(1)-P(3), 2.282(2); Co(1)-I(1), 2.488(1); Co(1)-B(1), 3.490(8). P(1)-Co(1)-P(2),
90.33(8); P(1)-Co(1)-P(3), 94.46(7); P(2)-Co(1)-P(3), 91.95(7); P(1)-Co(1)-I(1), 117.54(6); P(2)-Co(1)-I(1), 129.19(6); P(3)-Co(1)-I(1), 124.01(6).
(B) {[PhBP3]Co(µ-Br)}2, (2); Selected interatomic distances (Å) and angles (deg): Co(1)-Co(1)#1, 3.668(1); Co(1)-P(3), 2.235(2); Co(1)-P(2), 2.247(2);
Co(1)-P(1), 2.336(2); Co(1)-Br(1), 2.413(1); Co(1)-Br(2), 2.424(1); Co(1)-B(1), 3.570(6). Co(1)-Br(1)-Co(1)#1, 98.97(5); Co(1)-Br(2)-Co(1)#1,
98.37(5); P(3)-Co(1)-P(2), 87.42(6); P(3)-Co(1)-P(1), 92.35(6); P(2)-Co(1)-P(1), 93.44(6); P(3)-Co(1)-Br(1), 165.70(5); P(2)-Co(1)-Br(1), 93.63(5);
P(1)-Co(1)-Br(1), 101.81(4); P(3)-Co(1)-Br(2), 91.35(5); P(2)-Co(1)-Br(2), 153.37(5); P(1)-Co(1)-Br(2), 113.19(4); Br(1)-Co(1)-Br(2), 81.33(3).
Symmetry transformations used to generate equivalent atoms: #1 - x, y, -z + 1/2. (C) {[PhBP3]Co(µ-Cl)}2, (3); Selected interatomic distances (Å) and
angles (deg): Co(1)-Co(2), 3.497(1); Co(1)-B(1), 3.551(3); Co(2)-B(2), 3.598(3); Co(1)-P(1), 2.224(1); Co(1)-P(2), 2.244(1); Co(1)-P(3), 2.391(1);
Co(1)-Cl(1), 2.284(1); Co(1)-Cl(2), 2.321(1); Co(2)-P(4), 2.227(1); Co(2)-P(5), 2.239(1); Co(2)-P(6), 2.336(1); Co(2)-Cl(1), 2.314(1); Co(2)-Cl(2),
2.332(1). Co(1)-Cl(1)-Co(2), 99.02(3); Co(1)-Cl(2)-Co(2), 97.46(3); P(1)-Co(1)-P(2), 88.16(3); P(1)-Co(1)-Cl(1), 162.76(3); P(2)-Co(1)-Cl(1),
92.82(3); P(1)-Co(1)-Cl(2), 92.55(3); P(2)-Co(1)-Cl(2), 159.70(3); Cl(1)-Co(1)-Cl(2), 80.65(2); P(1)-Co(1)-P(3), 91.09(3); P(2)-Co(1)-P(3), 92.66(3);
Cl(1)-Co(1)-P(3), 106.04(3); Cl(2)-Co(1)-P(3), 107.60(3); Cl(1)-Co(2)-Cl(2), 79.79(2).
metathetically from the iodide complex: in situ iodide abstrac-
tion by TlPF6 in THF, followed by addition of KBr or NaCl,
converted [PhBP3]CoI to {[PhBP3]Co(µ-Br)}2, (2), and {[PhBP3]-
Co(µ-Cl)}2, (3), respectively. Complexes 2 and 3 were isolated
in yields typically greater than 80% when the metathesis protocol
was repeated twice prior to workup. The halide derivatives 1,
2, and 3 gave rise to distinct and well-resolved, though
paramagnetically shifted, 1H NMR spectra that are provided for
reference in the Supporting Information.
Displacement ellipsoid representations for the solid-state struc-
tures (collected at 98 K) of 1, 2, and 3 are depicted in Figure
1A, 1B, and 1C, respectively. The structure shown to the right
for each complex provides a simplified representation in which
the aryl carbon atoms from the diphenylphosphine donors have
been omitted. Furthermore, the structures on the right of the
bromide and chloride derivatives are rotated so that the bridging
halide ligands are eclipsed. Only the bridged halide protruding
out of the plane of the page is visible.
The iodide complex 1 is green whereas the chloride and
bromide complexes 2 and 3 are purple in the crystalline state.
This difference in color between polycrystalline samples of the
three complexes correlates with their solid-state structures, which
were determined by low-temperature X-ray diffraction studies.
Most striking in Figure 1 is the monomeric structure obtained
for the iodide complex 1 in comparison to the dimeric solid-
state structures obtained for the bromide and chloride complexes
2 and 3, respectively. Although none of the three complexes
adopts an idealized local geometry, they can be approximated
9
15338 J. AM. CHEM. SOC. VOL. 124, NO. 51, 2002

Low Spin Cobalt(II) System
A R T I C L E S
as structure types typical of cobalt(II). The iodide 1 is best
described as pseudotetrahedrally coordinated, with low formal
symmetry due to a strong axial distortion and inequivalent Co-
phosphine bond lengths and angles. By contrast, the bromide
and chloride derivatives can be regarded as square pyramidal
structures at the localized cobalt center of each respective
dimeric unit. In each of the three structures, the tridentate
[PhBP3] ligand exhibits two short and one modestly elongated
Co-P bond. The P-Co-P angles vary only slightly from 90°.
The bromide complex 2 features a rigorously planar Co2Br2
rhombus that is bisected by a crystallographic C2-axis running
through the two bromide ligands. The C2-axis effectively places
the elongated axial Co-P bonds on opposite faces of the Co2Br2
plane. By comparison, the Co2Cl2 unit of the chloride complex
3 gently buckles from planarity. The two elongated, axial Co-P
bonds in this case reside on the same face of the Co2Cl2 unit.
The gentle buckling in 3, along with the decreased size of the
chloride bridge, effectively slides the two cobalt centers closer
together by comparison to their distance in 2 (3.668(1) Å in 2
versus 3.497(1) Å in 3). Dimeric 3 is structurally very similar
to its dicationic relative [{(triphos)Co(µ-Cl)}2][BPh4]2: the Co-
Co distance in [{(triphos)Co(µ-Cl)}2][BPh4]2 is 3.52 Å, and its
Co-Cl-Co angle is 100.0°, to be compared with an average
Co-Cl-Co angle of 98.2° in 3.23c,24 The solid-state structure
of [{(triphos)Co(µ-OH)}2][BPh4]2 has also been determined and
is similarly dimeric, though the O-Co-O angles are much
smaller and average 69.5°.23c Spectroscopic data confirms that
the bromide derivative [{(triphos)Co(µ-Br)}2][BPh4]2 is dimeric
as well;23c its solid-state crystal structure has not yet been
reported.
The solid-state structure of the iodide complex 1 exhibits
approximate Cs symmetry.25 An appreciable axial distortion
gives rise to I-Co-P bond angles that are much larger than
those found in idealized tetrahedral structures (P(1)-Co(1)-
I(1), 117.54(6); P(2)-Co(1)-I(1), 129.19(6); P(3)-Co(1)-I(1),
124.01(6)). Complex 1 is gently distorted from molecular 3-fold
symmetry by a modest elongation of one of its phosphine donors
by comparison to the other two (Co(1)-P(1), 2.200(2); Co(1)-
P(2), 2.206(2); Co(1)-P(3), 2.282(2)). Worth noting is that
monomeric 1 is formally a 15 electron complex; this is very
unusual for cobalt(II) systems supported by three phosphine
donors. Similar 15 electron, monomeric and pseudotetrahedral
cobalt(II) halides supported by tris(pyrazolyl)- and tris(thio-
ether)borate ligands have been reported previously.26,27 An
attempted synthesis of the analogous “[(triphos)CoI]+” system
was reported to have failed due to spontaneous reduction of
the cobalt(II) center to cobalt(I).23c
Figure 2. (A) SQUID plot of µeff (BM) versus T: [PhBP3]CoI, 1, (]);
[PhB(CH2PPh2)(CH2P(O)Ph2)2]CoI, 4, (9); [PhBP3]Co(O-2,6-dimethyl-
phenyl), 5, (4). (B) SQUID plot of øm-1 (mol/cm3) versus T: [PhBP3]CoI,
1, (]); [PhB(CH2PPh2)(CH2P(O)Ph2)2]CoI, 4, (9); [PhBP3]Co(O-2,6-
dimethylphenyl), 5, (4).
fitted in the temperature range specified to use data that obeyed
the Curie-Weiss law reasonably well. This was established from
-1
øm versus T plots, which are shown in Figures 2b and 3b.
For each case, the samples studied by SQUID magnetometry
had provided satisfactory combustion analyses, as recorded in
the Experimental Section. These same sample batches were also
used to obtain the EPR spectra that are discussed in the
following section.
Figure 2a displays the temperature dependence of the
calculated magnetic moment µeff (see Experimental Section for
-1
details) for complex 1, and Figure 2b plots its øm versus
IIb. Magnetic Data (SQUID) for {[PhBP3]CoIIX} (X ) I,
Br, Cl). The temperature dependence of the magnetic suscep-
tibility for complexes 1, 2, and 3 was studied by SQUID
magnetometry. Average magnetic moments were adjusted for
diamagnetic contributions using Pascal’s constants and were
temperature. A øm versus temperature plot for 1 is provided in
the Supporting Information. The magnetic data for 1 is
unexpected for a 4-coordinate cobalt(II) system in a pseudo-
tetrahedral coordination geometry. Its magnetic moment shows
little variation in the temperature range between 30 and 220 K.
The Curie law observed in this range indicates that the doublet
state is the only state that is thermally populated. At very low
temperature, intermolecular exchange phenomena likely quench
the bulk paramagnetism (Θ ) -1.38 ( 1.27 K). Above 220
K, the effective moment increases, albeit very gradually,
possibly suggesting that a high spin state is slightly but
increasingly populated as the temperature rises above 220 K.
The concentration of a high spin state at the elevated temper-
(25) Complex 1 contained two crystallographically independent molecules in
the asymmetric unit. For brevity, only one of the molecules is explicitly
discussed in the main text. The Supporting Information contains crystal
coordinates, bond lengths, and bond angles for both molecules of the
asymmetric unit.
(26) Reinaud, O.; Rheingold, A.; Theopold, K. Inorg. Chem. 1994, 33, 2306-
2308.
(27) For a discussion of high spin [PhTttert-butyl]CoCl see: Schebler, P.; Riordan,
C.; Guzei, I.; Rheingold, A. Inorg. Chem. 1998, 37, 4754-4755. The
complex [PhTttert-butyl]CoI has also been fully characterized but is not yet
published (Riordan, C., personal communication).
9
J. AM. CHEM. SOC. VOL. 124, NO. 51, 2002 15339

A R T I C L E S
Jenkins et al.
atures is very small, however, as even at 300 K the value for
µeff only reaches 2.85 µB. We note that the gradual rise in the
moment that is observed above 220 K for 1 was reproducible
and does not appear to be an artifact of the experiment. An
average magnetic moment of 2.67 µB was obtained for 1 by
fitting the susceptibility data in the temperature range between
10 and 310 K. A fit to the region from 30 to 220 K, where the
sample shows good Curie-Weiss behavior, provides an average
moment of 2.65 ( 0.03 µB, the error representing two standard
deviations. For a 4-coordinate cobalt(II) complex, this moment
would typically imply a cobalt(II) ion in an approximate square
planar geometry.28 Similar low spin moments are also common
for cobalt(II) systems with higher coordination numbers (e.g.,
square pyramidal, trigonal bipyramidal, or octahedral type
structures),38 the value of 2.65 µB tending toward the high side
of what has been typically observed for low spin cobalt(II). By
contrast, magnetic moments for tetrahedral and pseudo-
tetrahedral structure types more closely related to [PhBP3]CoI
(Td, C3V, and lower symmetries) typically fall within a range
between 4.3 and 5.2 µB and are assigned as high spin.38,42 The
average g value of 3.06 that is extracted from the SQUID
magnetization data, assuming a moment of 2.65 µB, is much
higher than the effective g value directly obtained by solid-
state EPR spectroscopy (g ≈ 2.1), as discussed below. The crude
relationship between g and µeff, as defined by the simplified
equation µeff ) g{S(S + 1)}1/2, assumes complete quenching
of an orbital contribution,45 which may be a serious over-
simplification in the present context. The room-temperature
benzene solution magnetic moment of 1 was estimated by the
(28) Casey, A. T.; Mitra, S. Theory and Application of Molecular Para-
magnetism; Bourduaux, E. A., Mulay, N. L., Eds.; Wiley: New York, 1976.
(29) (a) Sur, S. K. J. Magn. Reson. 1989, 82, 169-173. (b) Evans, D. F. J.
Chem. Soc. 1959, 2003-2005.
(30) (a) Sinn, E. Coord. Chem. ReV. 1970, 5, 313-347, and references therein.
(b) Ball, P. W. Coord. Chem. ReV. 1969, 4, 361-383. (c) Kato, M.;
Jonassen, H. B.; Fanning, J. C. Chem. ReV. 1964, 64, 99-128.
(31) Ghilardi, C. A.; Mealli, C.; Midollini, S.; Nefedov, V. I.; Orlandini, A.;
Sacconi, L. Inorg. Chem. 1980, 19, 2454-2462.
(32) Note that the solution susceptibilities for 2 and 3 are calculated by the
Evans method assuming a monomeric formulation for each complex.
(33) Sealy, R.; Hyde, J. S.; Antholine, W. E. Modern Physical Methods in
Biochemistry; Neuberger, A., Van Deenen, L. L. M., Eds.; Elsevier: New
York, 1985; p 69.
Figure 3. (A) SQUID plot of µeff (BM) versus T: {[PhBP3]Co(µ-Br)}2, 2,
(9);{[PhBP3]Co(µ-Cl)}2, 3, (]). (B) SQUID plot of øm-1 (mol/cm3) versus
T: {[PhBP3]Co(µ-Br)}2, 2, (9);{[PhBP3]Co(µ-Cl)}2, 3, (]).
method of Evans and provided a value of 2.87 µB, assuming a
diamagnetic correction from Pascal’s constants, in close agree-
ment with that obtained from the SQUID data at 300 K.29 We
suggest that the solid-state and solution moments are most
consistent with a low spin assignment for 1, and that a low
spin ground state is very reasonable for a complex of this type
based upon the rationale provided in the discussion section.
Moreover, the low spin assignment is further corroborated by
the EPR data for 1.
The solid-state magnetic susceptibilities for polycrystalline
samples of the dimeric bromide {[PhBP3]Co(µ-Br)}2 and
chloride {[PhBP3]Co(µ-Cl)}2 were also measured by SQUID
magnetometry and are plotted, per dimeric unit, in Figure 3,
parts A and B. The magnetic data obtained for the dimers is
more complex than the data obtained for monomeric 1. Each
set of susceptibility data was collected twice to verify its
experimental reproducibility. Little, if any, evidence for detect-
(34) Superhyperfine coupling to phosphorous in a 5-coordinate cobalt(II)
phosphine complex has been treated previously. See: Stelzer, O.; Sheldrick,
W. S.; Subramanian, J. J. Chem. Soc., Dalton Trans. 1976, 966-970.
(35) Gatteschi, D. J. Mol. Catal. 1985, 23, 145-150.
(36) Oxygenation of related (triphos)Co2+ complexes did not produce a similarly
isolable species. See: Heinze, K.; Huttner, G.; Zsolnai, L. Chem. Ber. 1997,
130, 1393-1403.
(37) Banci, L.; Bencini, A.; Benelli, C.; Gatteschi, D.; Zanchini, C. Struct.
Bonding 1982, 52, 37-86.
(38) Cotton, F. A.; Wilkinson, G. AdVanced Inorganic Chemistry, 5th ed.;
Wiley: New York, 1988; p 729. See ref 22 for a more thorough discussion
and representative references.
(39) Osmometric measurements to determine the solution molecular weights of
2 and 3 by the method of Signer were vitiated due to the tendency of these
species to precipitate from toluene and benzene solution at moderate
concentrations.
(40) Sacconi, L.; Ciampolini, M.; Speroni, C. P. Inorg. Chem. 1965, 4, 1116-
1119.
(41) Nicolini, M.; Pecile, C.; Turco, A. J. Am. Chem. Soc. 1965, 87, 2379-
2384.
able antiferromagnetic coupling in dimers 2 and 3 can be
(42) For relevant references that deal with 4-coordinate cobalt(II) in approximate
C3 or C3V symmetry, see the following: (a) Garrett, B. B.; Goedken, V.
L.; Quagliano, J. V. J. Am. Chem. Soc. 1970, 92, 489-493. (b) Bertini, I.;
Gatteschi, D.; Mani, F. Inorg. Chim. Acta 1973, 7, 717-720. (c) Gerloch,
M.; Hanton, L. R. Inorg. Chem. 1980, 19, 1692-1698. (d) Gerloch, M.;
Manning, M. R. Inorg. Chem. 1981, 20, 1051-1056. (e) Banci, L.; Benelli,
C.; Gatteschi, D.; Mani, F. Inorg. Chem. 1982, 21, 1133-1136.
(43) A series of related 5-coordinate [PhBP3]Co(II) complexes has been prepared
that will be reported in due course.
-1
gleaned from the temperature-dependent plots of µeff and øm
.
The acid test for antiferromagnetic exchange coupling in a
1
dimeric complex consisting of two spin /2 ions is that the
susceptibility should reach a maximum as the temperature of
the sample is lowered, but then fall precipitously to zero as the
sample is cooled further.45 For both 2 and 3, the magnetic
moments decrease very gently from 310 K down to about 80
K, at which point they level off until reaching the lowest
(44) Jazdzewski, B. A.; Holland, P. L.; Pink, M.; Young, V. G.; Spencer, D. J.
E.; Tolman, W. B. Inorg. Chem. 2001, 40, 6097-6107.
(45) Kahn, O. Molecular Magnetism; VCH Publishers: New York, 1993; pp
1-10.
9
15340 J. AM. CHEM. SOC. VOL. 124, NO. 51, 2002

Low Spin Cobalt(II) System
A R T I C L E S
temperature region where intermolecular paramagnetic quench-
ing begins to occur. The expected maximum in the susceptibility
for 2 and 3 is not observed, in contrast to the temperature-
dependent susceptibilities Sacconi reported for [{(triphos)Co-
(µ-X)}2][BPh4]2 (X ) Br, Cl, OH), indicating the dicationic
dimers exhibit strong antiferromagnetic coupling (J < -100
cm-1) between their Co2+ centers.23c Antiferromagnetic ex-
change behavior is most typical of halide-bridged cobalt(II),
copper(II), and titanium(I) dimers, where an unpaired electron
resides at each metal center,30 and it is emphasized that
antiferromagnetic exchange does not appear to be present in
either 2 or 3. The net change in the moment for both 2 and 3
is rather small between 300 and 80 K. For example, at 300 K
the moment of chloride 3 is 2.9 µB, which decreases to 2.5 µB
at 80 K. For bromide 2, the moment at 300 K (2.3 µB) is
appreciably lower than for 3 and only decreases to a value of
1.8 µB at 80 K. For a simple ferromagnetic exchange interaction
(J > 0) between the two spin 1/2 ions in dimers 2 and 3, a plot
of ømT versus T is expected to show a rise in the curve as the
temperature is lowered. Instead, we observe a gentle decrease
in ømT (and likewise µeff) as the temperature is lowered, which
decreases dramatically only at very low temperature due to the
onset of intermolecular magnetic quenching interactions. The
ømT versus T plots can be found in the Supporting Information
for 2 and 3. A fit of the lower temperature data, where the
Curie-Weiss law is obeyed reasonably well (T < 110 K),
provided a magnetic moment µeff ) 1.77 ( 0.08 µB and Θ )
-1.48 ( 0.49 K for complex 2. A moment of µeff ) 2.47 (
0.14 µB and theta value of Θ ) -3.13 ( 0.67 K was obtained
for complex 3 when fit in the same temperature regime. Zero-
field splitting, weak exchange, and mixing of the triplet and
singlet states of 2 and 3 likely complicate the observed
magnetism and contribute to the gradual attenuation rather than
rise of the moment of each sample as the temperature is lowered,
as would have been expected for systems exhibiting pronounced
ferromagnetic exchange. Regardless, we suggest that the
magnetic data for 2 and 3 are most consistent with assigning
them to S ) 1 ground states. The EPR data for solid samples
of 2 and 3 are also most consistent with a triplet ground state
assignment (see below); the signal amplitude for these dimers
increases as the temperature of each sample is lowered from
25 to 3.6 K, as is to be expected for a weak ferromagnetically
Figure 4. (A) EPR spectrum of a glassy toluene solution of [PhBP3]CoI
at 17 K. Instrumental parameters: ν ) 9.477 GHz, modulation frequency
) 100 kHz, modulation amplitude ) 4 G, microwave power ) 0.51 mW,
conversion time ) 81.92 ms, time constant ) 20.48 ms, 3 scans. g ≈ 2
region and spectral assignment; g and A were estimated by simulating
||
||
the EPR spectrum (inset). (B) EPR spectrum of a glassy toluene solution
of [PhBP3]CoI in toluene-d8 at 15 K. Instrumental parameters: ν ) 9.476
GHz, modulation frequency ) 100 kHz, modulation amplitude ) 3 G,
microwave power ) 0.202 mW, conversion time ) 81.92 ms, time constant
) 20.48 ms, sweep time ) 83.9 s, 3 scans.
K, is not so pronounced in solution. Evans method measure-
ments for 2 and 3 in benzene-d6 at 25 °C provided moments of
2.7 µB and 3.0 µB, respectively, rather similar to the solution
moment obtained for 1.32 As will be suggested by the EPR data
discussed in the following section, these solution moments
reflect a change in both geometry and spin state for 2 and 3 in
toluene solution.
IIc. EPR Spectra for {[PhBP3]CoIIX} (X ) I, Br, Cl). The
EPR spectra for 1, 2, and 3 are extremely sensitive to oxygen
due to a high spin impurity that results from a well-defined
ligand oxidation process that has been carefully studied for the
case of 1. High-quality spectra for all three complexes were
only acquired using thoroughly recrystallized samples that
afforded satisfactory combustion analysis. The absence of high
spin oxidation products was further confirmed by optical
spectroscopy prior to EPR data collection.
The EPR spectrum in glassy toluene (17 K) of a pure sample
of 1 is shown in Figure 4A. Due to the high oxygen sensitivity
of 1, the spectra that were originally obtained showed a rather
broad feature in the region g ≈ 4.8 (ca. 120 mT). This feature
was entirely removed upon rigorous exclusion of oxygen from
the sample. The broad feature at ∼4.8 corresponds to a signal
1
coupled dimer with two S ) /2 spin centers.
A dimeric cobalt(II) complex whose magnetism may be
related to that observed for 2 and 3 was provided by Sacconi
and co-workers in 1980.31 They reported that the µ-thiolato
complex, [{(triphos)Co(µ-SCH3)}2][BPh4]2, provided a µeff of
1.8 µB that gradually decreased to 0.8 µB at 130 K, at which
point it leveled off down to the lowest temperature for which
data was recorded (87 K). The reasoning Sacconi and co-
workers provided to explain the odd magnetic behavior of
[{(triphos)Co(µ-SCH3)}2][BPh4]2 was that antiferromagnetic
1
exchange occurred between the spin /2 cobalt centers through
the µ-SCH3 units, but that the limiting value represented a
contribution from partial occupation of a higher triplet state. In
our case, the limiting value at lower temperature represents a
triplet state likely mixed with a singlet state that is close in
energy.
It should be noted that the difference between the total µeff
for solid samples of 2 and 3, a difference of ca. 0.6 µB at 300
9
J. AM. CHEM. SOC. VOL. 124, NO. 51, 2002 15341

A R T I C L E S
Jenkins et al.
Scheme 2
from the high spin complex 4, whose independent characteriza-
tion will be discussed shortly. The spectrum of pure 1 is nearly
independent of temperature (except for its intensity) in the
temperature range 20 K to 140 K. It is characterized by an axial
this model, we could achieve only qualitative agreement between
the simulated and experimental spectra. On the basis of these
simulations, a zero-field splitting of roughly 0.2 cm-1 or less is
consistent with the observed EPR spectra. Also, we note that
the EPR spectrum of a monomeric triplet complex, [PhBP3]-
Co(PMe3), has been recorded and contains features qualitatively
similar to those of 2 and 3 in the solid-state, including the low
field g signal ∼4.8. This system has been briefly described
elsewhere.18d
g factor (g ) 2.20 and g ) 2.05), and poorly resolved 59Co
||
(I ) 7/2) hyperfine splitting. However, a very complex super-
hyperfine splitting pattern was resolved below 50 K in toluene-
d8 (Figure 4B),33 ascribable to the 31P (I ) 1/2)34 and possibly
127I- (I ) 5/2) donor atoms. An isotropic EPR spectrum with
giso ) 2.1 was observable at room temperature, further confirm-
ing the low spin nature of 1. The powder EPR spectrum of 1
shows features analogous to those of the frozen toluene sample,
and a representative solid-state spectrum is provided in the
Supporting Information. EPR signals for high spin cobalt(II)
systems are typically observed only below ∼30 K due to the
fast spin-lattice relaxation times of the high spin Co(II)
nucleus.48 That we are able to observe an intense signal for 1
even at room-temperature further supports the notion that 1 has
a doublet ground-state both in solution and in the solid-state.
The solid-state magnetic data that were presented above for
2 and 3 suggested that they each have triplet ground states close
in energy to possibly mixed singlet excited states, and that the
magnitude of exchange energy J is difficult to measure from
the SQUID data alone. To the extent that ferromagnetic
exchange is present, it is likely very small (<10 cm-1). To
further explore each of these systems, their EPR spectra were
recorded using the same samples that had been carefully purified
for the magnetic measurements. As for 1, admission of adventi-
tious oxygen to samples of 2 and 3 produced oxidation products
(vide infra) that afforded spectra with signals centered at g ≈
4.8. Problematic is that these features coincidentally overlap
with the low field EPR signal expected (and observed in the
solid-state) for the pure dimers of 2 and 3. Again, rigorous
exclusion of oxygen solved this problem and high quality spectra
were obtained. The EPR spectra (Figure 5A, 6A) for pure
polycrystalline samples of both 2 and 3 are well resolved below
30 K. Each spectrum supports the suggestion of the triplet
species at low temperature: the intensity of these spectra were
not attenuated, but rather increased, from 25 to 3.6 K, consistent
with weak ferromagnetic coupling. The temperature dependence
of the solid-state EPR spectrum of 2 is shown in Figure 5d.
We attempted to simulate the EPR spectra of the polycrystalline
samples of 2 and 3 under the assumption that the exchange
energy |J| . than the X-band microwave quantum (0.3 cm-1).
In such cases the triplet state can be described in terms of D
and E using the standard spin-Hamiltonian for an S ) 1 spin
system (see the Experimental Section).48 We also made the crude
assumption that the dipolar and g tensors were collinear.35 Using
Dissolution of purple crystals of {[PhBP3]Co(µ-Br)}2 (2) and
{[PhBP3]Co(µ-Cl)}2 (3) in toluene was accompanied by a drastic
color change to bright green, similar to the color of [PhBP3]-
CoI both in solution and in the crystalline state. This behavior
suggested to us a structural change in the dimers 2 and 3 upon
dissolution. Indeed, we found by EPR that in solution a
temperature-dependent equilibrium between monomeric and
dimeric species seems to exist (Scheme 2). The EPR spectrum
of 2 in glassy toluene (15 K) is shown in Figure 5b. This
spectrum presumably represents a monomeric species akin to
the structurally characterized iodide monomer, perhaps with an
additional trace presence of the dimeric form of 2. The
observation that g
>
for 2 in the frozen glass clearly
||
g
establishes that the square-pyramidal arrangement of the cobalt
complex, observed in the solid-state, is not preserved in
solution.21 We suggest that the geometry of 2 in solution is that
of a pseudotetrahedral, 4-coordinate monomer, as for the case
of 1. The similarity of the magnetic moments for solutions of
1 and 2 is also consistent with their having similar coordination
geometries. We therefore conclude that the glassy toluene EPR
spectrum of 2 represents the monomeric and low spin complex
[PhBP3]CoBr. The resolved hyperfine coupling in the g region
||
is in accord with this model (A ) 88 × 10-4 cm-1) as it shows
||
an octet, indicative of coupling to a single cobalt nucleus. If
the dimer of 2 were maintained in solution, we would have
expected to observe a 15-line splitting pattern. A related EPR
study was also performed for the chloride 3. The polycrystalline
and glassy EPR spectra of 3, shown in Figure 6, parts a and b,
respectively, are similar to that of 2. The major difference is
that a readily detectable population of the dimer of 3 appears
to be present in glassy toluene at 12 K. The low spin monomer
[PhBP3]CoCl, however, is the major species present in the glass
even at this low temperature. The additional hyperfine reso-
nances observed in the g region (∼2.2) for 3 are also suggestive
||
of a modest concentration of the triplet {[PhBP3]Co(µ-Cl)}2 in
the glassy toluene spectrum, that is super-imposed with the
signal expected for [PhBP3]CoCl. The hyperfine coupling for
[PhBP3]CoCl is resolvable with A ) 92 × 10-4 cm-1. The
||
temperature dependence of the optical spectrum of 3 is discussed
below and lends further support to the suggestion that 3 can
dimerize in solution as a function of temperature.
(46) Swalen, J. D.; Gladney, H. M. IBM J. Res. DeV. 1964, 8, 515-526.
(47) van Veen, G. J. Magn. Resonance 1978, 30, 91-109.
(48) (a) Pilbrow, J. R. Transition Ion Electron Paramagnetic Resonance;
Clarendon Press: Oxford, 1990. (b) Aasa, R.; Va¨nngård, T. J. Magn.
Resonance 1975, 19, 308-315.
IId. Conversion of Low Spin Iodide, 1, to the High Spin
Complexes [PhB(CH2P(O)Ph2)2(CH2PPh2)]CoI and [PhBP3]-
9
15342 J. AM. CHEM. SOC. VOL. 124, NO. 51, 2002

Low Spin Cobalt(II) System
A R T I C L E S
Figure 5. (A) EPR spectrum of a polycrystalline sample of {[PhBP3]Co(µ-Br)}2 (15 K). Instrumental parameters: ν ) 9.480 GHz, modulation frequency
) 100 kHz, modulation amplitude ) 4 G, microwave power ) 1.01 mW, conversion time ) 163.84 ms, time constant ) 40.96 ms, 10 scans. (B) EPR
spectrum of a glassy toluene solution of {[PhBP3]Co(µ-Br)}2 at 12 K. Instrumental parameters: V ) 9.472 GHz, modulation frequency ) 100 kHz, modulation
amplitude ) 4 G, microwave power ) 0.101 mW, 1 scan. (C) EPR spectrum of a glassy toluene solution of {[PhBP3]Co(µ-Br)}2 at 50 K. g and A were
||
||
estimated by simulating the EPR spectrum. Instrumental parameters: V ) 9.477 GHz, modulation frequency ) 100 kHz, modulation amplitude ) 5 G,
microwave power ) 2.02 mW, conversion time ) 20.48 ms, time constant ) 5.12 ms, 20 scans. (D) Series of EPR spectra for a polycrystalline sample of
{[PhBP3]Co(µ-Br)}2 plotted to show the temperature dependence of the EPR signal. The signal amplitude increases as the temperature decreases, suggestive
of a ferromagnetic triplet ground state.
Scheme 3
Co(O-2,6-Me2-C6H4). It was prudent to prepare one or more
complexes of similar geometry to 1 with a high spin ground
state so that the relevant structural, magnetic, and spectroscopic
data could be compared to the monomeric low spin derivatives
already described.
The oxygen sensitivity of the halides 1, 2, and 3 encouraged
us to assess the stoichiometric reaction between 1 and molecular
oxygen. We were fortunate to find that the addition of either
stoichiometric or excess dioxygen to a benzene solution of 1
resulted in a rapid color change from green to bright blue. The
product, which formed quantitatively according to optical and
1H NMR spectroscopies, proved to be a 4-coordinate cobalt
iodide complex whose tripodal borate ligand had undergone a
4-electron oxidation process (Scheme 3).36 The solid-state
structure of the product, [PhB(CH2P(O)Ph2)2(CH2PPh2)]CoI, (4),
shown in Figure 7A, has approximate Cs symmetry with a mirror
plane bisecting the two phosphine oxide donors. By comparison
to its green precursor 1, blue 4 has several structural features
worth highlighting. The Co-I and Co-P bond lengths are
appreciably elongated in 4 by comparison to 1, qualitatively
consistent with an expanded high spin cobalt radius (vide infra).
Furthermore, the borate counteranion of the oxidized ligand is
tethered further away from the cobalt center in 4 (Co-B distance
is 3.77 Å in 4 and 3.49 Å in 1). This fact, coupled with a
9
J. AM. CHEM. SOC. VOL. 124, NO. 51, 2002 15343

A R T I C L E S
Jenkins et al.
We also prepared a more conventional, 4-coordinate and high-
spin complex supported by an intact [PhBP3] ligand. The iodide
complex 1 was allowed to react with the moderately bulky
thallium aryloxide reagent Tl(O-2,6-Me2Ph), which smoothly
generated the aryloxide species [PhBP3]Co(O-2,6-Me2Ph), (5),
as a red-brown, crystalline solid (Scheme 3). The solid-state
structure of 5, depicted in Figure 7b, reveals 5 to be monomeric,
but grossly distorted from idealized 3-fold symmetry. Each of
the three O-Co-P angles is quite different (134.9(1)°, 120.4(1)°,
and 105.3(1)°) and one of the phosphine donor arms is slightly
elongated by comparison to the other two (2.366(1) Å versus
2.332(1) and 2.343(1) Å, respectively). High spin, 4-coordinate
cobalt(II) systems typically adopt higher symmetry when
possible; the distorted structure obtained for 5 likely results from
steric interactions between the bulky aryloxide and [PhBP3]
ligands. The Co-P bond lengths in 5 are, on average, ap-
preciably elongated in comparison to 1 (∼0.12 Å). This
difference again qualitatively reflects their different spin states.
IIe. Comparative Magnetic and EPR Characterization of
[PhB(CH2P(O)Ph2)2(CH2PPh2)]CoI (4) and [PhBP3]Co(O-
2,6-Me2-C6H4) (5). The SQUID data obtained for blue 4, shown
in Figure 2, parts a and b, establish it to be a typical high spin
cobalt(II) complex. The average magnetic moment of
4.29 µB, obtained from the solid-state SQUID data fit between
10 and 300 K, falls within the range of the many 4-coordinate,
non square planar high spin systems that have been previously
described.38 A solution moment of 4.5 µB (C6D6, 25 °C) was
estimated by the method of Evans for 4. The EPR spectrum of
4 also reveals its formal change in spin state upon ligand
oxidation. Both low-temperature powder (Figure 8) and glassy
toluene samples of 4 gave rise to similar EPR spectra that
were distinct from those obtained for 1 and similar to those
normally observed for high spin cobalt(II) systems.37 The intense
signal (g ≈ 4.8) characteristic of 4 was also observed when 2
and 3 were exposed to increasing amounts of oxygen (see
Supporting Information), suggesting that similar products are
formed in each of the three cases. It is emphasized that the EPR
spectrum of 4 was only observed at low temperature (<50 K),
typical of high spin cobalt(II) derivatives and contrasting the
Figure 6. (A) EPR spectrum of a polycrystalline sample of {[PhBP3]Co-
(µ-Cl)}2 at 3.6 K. Instrumental parameters: ν ) 9.477 GHz, modulation
frequency ) 100 kHz, modulation amplitude ) 4 G, microwave power )
0.101 mW, conversion time ) 163.84 ms, time constant ) 40.96 ms, 4
scans. (B) EPR spectrum of a glassy toluene solution of {[PhBP3]Co(µ-
Cl)}2 at 12 K. Instrumental parameters: V ) 9.476 GHz, modulation
frequency ) 100 kHz, modulation amplitude ) 4 G, microwave power )
0.202 mW, conversion time ) 163.84 ms, time constant ) 40.96 ms, 7
scans. The inset g ≈ 2 region showing resolved cobalt hyperfine structure;
g
||
and A were estimated by simulating the EPR spectrum.
||
decrease in the donor strength of the [PhB(CH2P(O)Ph2)2(CH2-
PPh2)] ligand by comparison to its non-oxidized [PhBP3]
precursor, suggests that the cobalt center in 4 should experience
a weaker ligand-field than its precursor complex 1. The bromide
and chloride complexes appear to undergo an analogous ligand
oxidation reaction on exposure to oxygen, though the products
were not characterized crystallographically. Green solutions of
2 and 3 rapidly turned blue, and their optical spectra changed
accordingly. As alluded to above, partial oxidation of samples
of 1, 2, and 3 is readily discerned by EPR spectroscopy.
Figure 7. Displacement ellipsoid representation (50%): (A) [PhB(CH2PPh2)(CH2P(O)Ph2)2]CoI, (4); Selected interatomic distances (Å) and angles (deg):
Co-P(3), 2.335(1); Co-I, 2.566(1); Co-O(1), 1.937(2); Co-O(2), 1.937(2); Co-B, 3.768(4); O(1)-P(1), 1.513(2); O(2)-P(2), 1.514(2); O(1)-Co-O(2),
103.3(1); O(1)-Co-I, 117.25(7); O(1)-Co-P(3), 105.90(7); O(2)-Co-I, 115.55(7); P(3)-Co-I, 113.83(3); O(2)-Co-P(3), 98.85(7). (B) [PhBP3]Co(O-
2,6-dimethylphenyl), (5). Selected interatomic distances (Å) and angles (deg): Co-O, 1.851(2); Co-P(3), 2.332(1); Co-P(2), 2.366(1); Co-P(1), 2.343(1);
Co-B, 3.446(3); O-Co-P(3), 120.39(5); O-Co-P(2), 134.93(5); P(3)-Co-P(2), 98.27(2); O-Co-P(1), 105.27(5); P(3)-Co-P(1), 92.72(2); P(2)-
Co-P(1), 94.22(2).
9
15344 J. AM. CHEM. SOC. VOL. 124, NO. 51, 2002

Low Spin Cobalt(II) System
A R T I C L E S
Figure 8. Frozen solution EPR spectrum of [PhB(CH2P(O)Ph2)2(CH2PPh2)]-
CoI (toluene, 4 K). Instrumental parameters: ν ) 9.477 GHz, modulation
frequency ) 100 kHz, modulation amplitude ) 10 G, microwave power )
2.02 mW, conversion time ) 40.96 ms, time constant ) 10.24 ms, 5 scans.
spectrum of 1, where a strong isotropic signal was detected even
at room temperature. The aryloxide complex 5 also appears to
be a high spin species in the solid-state, as evident from its
SQUID data provided in Figure 2. Its high spin moment was
maintained in solution (4.3 µB, C6D6, 25 °C). It is also interesting
to note that the ømT versus T plot for 5 (see Supporting
Information) reveals an unexpected rise in ømT as the sample
temperature is lowered. This rise becomes more dramatic at
temperatures < 100 K and reaches a maximum at 16 K, at which
point it falls precipitously to zero as the temperature is lowered
further, perhaps suggestive of intermolecular antiferromagnetic
exchange. The glassy toluene EPR spectrum of 5 is also
provided in the Supporting Information. Its EPR signal is not
detected at elevated temperatures (>50 K).
IIf. Comparative Optical Spectra of [PhBP3]CoI (1),
{[PhBP3]CoBr}2 (2), {[PhBP3]CoCl}2 (3), and [PhB(CH2P-
(O)Ph2)2(CH2PPh2)]CoI (4). The room-temperature optical
spectra for green solutions of 1, 2, and 3 (250-1500 nm), and
for the high spin complex 4, were obtained in toluene and are
shown in Figure 9. The spectra for the halides 1, 2, and 3 are
very similar and are dominated by an apparent broad charge-
transfer band in the visible region that displays an expected blue
shift from I- (638 nm) to Br- (612 nm) to Cl- (594 nm). Lower
energy features that show less halide dependence are also present
in the spectrum of each of the three halides. It is important to
note that incubation of green toluene solutions of 1 and 2 at
-78 °C did not effect an appreciable color change. In contrast,
incubation of a green solution of chloride 3 at -78 °C gave
rise to a distinctly purple solution whose green color was
recovered on warming. The inset spectrum shown in Figure 9
was acquired upon rapid removal of the purple solution
containing 3 from a -78 °C cold bath. In accord with the EPR
data, we presume that this spectrum represents a superposition
of the spectrum for the dimeric form of 3 and its monomeric,
pseudo-tetrahedral low spin form. The optical spectrum of high
spin 4 is consistent with related tetrahedral and distorted
tetrahedral cobalt(II) complexes.38 The visible region of the
spectrum of 4 displays moderately intense bands in the range
580 to 700 nm. A lower energy, broad band is observed ∼1020
nm. Theopold reported similar features for the optical spectrum
of a blue, tris(pyrazolyl)borate-supported cobalt(II) iodide
monomer in a distorted tetrahedral geometry.26 The optical
spectrum of high spin 5 was also obtained in toluene solution.
The visible region of its spectrum was characterized by broad
Figure 9. Absorption spectra for compounds [PhBP3]CoI, 1, (‚‚‚), {[PhBP ]-
Co(µ-Br)}2, 2, (- - -), {[PhBP3]Co(µ-Cl)}2, 3, ()), [PhB(CH2PPh23)-
(CH2P(O)Ph2)2]CoI, 4, (b) between 500 and 1500 nm. The similarity
between each of the spectra suggests that each complex is predominantly
monomeric at ambient temperature in toluene solution. Inset: {[PhBP3]-
Co(µ-Cl)}2, 3, at 298 K ()) and at 220 K (9) in toluene solution, illustrating
the temperature dependence of its optical absorption that is associated with
a monomer (298 K) to dimer (220 K) equilibrium.
features between 400 and 800 nm with resolvable maxima at
530, 580, and 752 nm.
III. Discussion
The preceding sections have presented structural, magnetic,
and spectroscopic evidence to suggest that a low spin formula-
tion is correct for monomeric [PhBP3]CoX systems, where X
represents the halides I-, Br-, or Cl-. Although the iodide
complex 1 is a simple, 4-coordinate monomer both in the solid-
state and in solution, the chloride 3, and to a lesser extent the
bromide 2, are capable of dimerizing to a modest degree in
solution at low temperature; they both crystallize in the dimeric
form. The solid-state SQUID and EPR data collected for the
dimers 2 and 3 prompt us to assign them to triplet ground states
exhibiting weak ferromagnetic coupling. A more detailed
analysis of their electronic structures is worth pursuing in future
studies, though it is at present clear that the electronic structures
of 2 and 3 are quite distinct from the structurally related dimers
of Sacconi, these latter systems being characterized by singlet
ground states due to pronounced antiferromagnetic exchange.15
At room temperature, both 2 and 3 are predominantly
monomeric in solution.39 Room-temperature solutions of 1, 2,
and 3 in toluene are bright green; solutions of 1 and 2 remain
green on cooling to 195 K, whereas a solution of 3 turns purple
(reversibly) as it is cooled, indicative of the dimer concentration
at low temperature. The solution EPR spectra are also consistent
with this model. At low temperatures (<50 K), the EPR spectra
of iodide 1 and bromide 2 show only the monomer form. For
the chloride 3, however, the dimeric form also appears to be
present below 50 K, superimposed with the spectrum for the
monomer. The optical spectra for the three complexes are in
accord with these data. Each of the spectra shows very similar
features at room temperature, suggesting analogous stereochem-
ical environments in solution. Cooling the solution of 3 causes
a dramatic change in color from green to purple, reflecting the
increase in concentration of its dimeric form.
9
J. AM. CHEM. SOC. VOL. 124, NO. 51, 2002 15345

A R T I C L E S
Jenkins et al.
The most intriguing observation concerns the low spin
character of the monomeric form of these simple halide
complexes. Configurational solution equilibria between high and
low spin complexes of cobalt were first observed more than
three decades ago.8,40,41 The typical situation is as follows: a
low spin pentacoordinate cobalt(II) complex can dissociate one
ligand to afford a (i) high spin, tetrahedral or distorted tetrahedral
structure type, or (ii) a low spin, square planar or distorted square
planar structure type. In certain cases, a solution equilibrium
has been observed between strictly quadri-coordinate structures
that are high spin (tetrahedral) and low spin (square planar),
such as the bis(â-ketoamino)cobalt(II) complexes described by
Holm and Everett.8a Stereochemical tuning of a series of
tropocoronand ligands coordinated to divalent cobalt by Lippard
and co-workers was shown to give rise to a range of intermediate
structure types that were classified as either distorted tetrahedral
(and high spin) or distorted square planar (and low spin).9 The
halides described in the present manuscript are anomalous in
that, for complexes 2 and 3, solution equilibria between a low
spin, 5-coordinate, distorted square pyramidal structure and a
low spin, 4-coordinate, distorted tetrahedral structure apparently
exist. The X-ray crystal structure of the iodide complex 1
provides a structural snapshot of the distorted tetrahedral, low
spin 4-coordinate structure, a structure that is not directly
observable for complexes 2 and 3, but which is inferred from
the solution data available for these complexes. It becomes
apparent that a low spin state for quadri-coordinate cobalt(II)
can be achieved by a distortion strikingly different from the
severe distortion that takes a tetrahedral structure to a square
planar structure. In the present case, a more subtle distortion,
one that is distinctly axial in character, gives rise to the low
spin configuration. The following discussion, accompanied by
Figures 10 and 11, provides a simple orbital explanation for
this phenomenon.
Figure 10. Qualitative d-orbital splitting diagram that depicts a descent in
symmetry from Td to C3V symmetry for an axially distorted tetrahedral
cobalt(II) system. As defined, the z-axis proceeds through one Co-L vector
of the starting tetrahedron. The figure presents three limiting structure
types: an idealized tetrahedron (CoL4), a C3V structure (CoYL3) with bond
angles analogous to the tetrahedral structure, and a C3V structure resulting
from a strong axial distortion (CoXP3) in which the P-Co-P angle are
fixed to ∼90 °. Note that this descent proceeds naturally from an octahedral
CoL6 species but that the correlation diagram places the z-axis at the
center of one face of the octahedron rather than its typical position directly
along one of the L-Co-L axes. The atomic compositions under Oh
symmetry for the t2g and eg set under this axis definition are as follows:49
For t2g: z2; {(2/3)1/2(x2 - y2) - (1/3)1/2yz}; {(2/3)1/2xy - (1/3)1/2xz}.
For eg: {(1/3)1/2x2 - y2 + (2/3)1/2yz}; {(1/2)1/2xy + (2/3)1/2xz}. The lower
portion of Figure 10 extends the splitting diagram to Cs symmetry by a
Jahn-Teller distortion that elongates one donor ligand and places the single
unpaired electron in a nondegenerate orbital stabilized by a lowering in σ*
character of one of the orbitals from the formally degenerate σ* e-set.
A tetrahedral ligand-field for divalent cobalt places three
degenerate orbitals, the t set, at a significantly higher energy
than the nonbonding e set. Following Hund’s rule, four electrons
fill the e set, and three electrons fill the t set, one per orbital in
a spin-aligned fashion. As is well-known, this is the origin of
the typically observed quartet ground state that has been
previously observed for tetrahedral and distorted tetrahedral
cobalt(II) complexes. Breaking the symmetry, whether to C3,
Cs, or C1, does not typically change this situation.42 The three
orbitals at higher energy are only modestly split as the symmetry
of the molecule descends from the tetrahedron. This splitting
of the degenerate t set is typically small by comparison to the
pairing energy that would be required to achieve a low spin
state. However, if this splitting becomes large, and if an orbital
of a1 symmetry (predominantly z2) from the upper set is strongly
stabilized on distortion, it should be possible to access a low-
spin population. We think that the stereochemical environment
and the strong ligand-field-donor-strength provided by the
[PhBP3] anion achieves this extreme and thereby provides access
to the low spin ground state for pseudotetrahedral cobalt(II).
The phosphine ligands of [PhBP3] are good sigma-donors and
is axially distorted along the Co-I vector (i.e., along z). When
stabilization of this a1 orbital becomes strong enough, and when
the ligand-field is sufficiently strong so as to destabilize the
now high-lying e-set, electron pairing becomes energetically
favorable and a low spin ground state is achieved. Figure 10
summarizes this model schematically by correlating the low spin
configuration in idealized C3V symmetry to idealized Td and Oh
structures. Note that the natural choice of coordinate axes for
all of the structures shown in Figure 10 places the z-axis through
one face of an octahedral ML6. Although this definition of
axes is obvious for C3V symmetry, it is not the typical choice
for Oh ML6. To describe the orbital character of the lower
t2g and upper eg sets, the orbital functions need to be correlated
to the new coordinate axis system. For an Oh ML6 complex,
these functions transform as:49 t2g set: z2; {(2/3)1/2(x2 - y2) -
(1/3)1/2yz}; {(2/3)1/2xy - (1/3)1/2xz}. For eg set: {(1/3)1/2x2 -
y2 + (2/3)1/2yz}; {(1/2)1/2xy + (2/3)1/2xz}. When the axial
should contribute to a strong ligand-field in pseudo-C3
1
whereby the upper orbital set, σ* in character, lies well above
the lower nonbonding orbital set. The dominant factor to
consider is the extent to which a hybrid a1 orbital (predominantly
z2 with some admixed s and pz) is stabilized as the cobalt center
(49) Orgel, L. E. An Introduction to Transition-Metal Chemistry; Wiley: New
York, 1960; p 174.
9
15346 J. AM. CHEM. SOC. VOL. 124, NO. 51, 2002

Low Spin Cobalt(II) System
A R T I C L E S
b). The electronic structure of 1 determined by DFT suggests
the HOMO orbital to be one of π-antibonding character between
the cobalt center and the iodide ligand, and the HOMO-1 orbital
to be predominantly dz2 in character, in accord with the
qualitative MO scheme presented in Figure 10.
Although a great many high spin distorted tetrahedral com-
plexes are known, two systems that deserve to be singled out
with respect to the present discussion are the monomeric halides
[Tp′′]CoI ([Tp′′] ) hydrotris(3-tert-butyl-5-methylpyrazolyl)-
borate) and [PhTttert-butyl]CoX ([PhTttert-butyl] ) phenyl(tris(tert-
butylthio)methyl)borate, X ) Cl- or I-), each of which has been
structurally characterized and assigned as high spin.26,27 Like
complex 1, they each have approximate C3 symmetry, and like
complex 1, they have a borate atom placed in close proximity
to the cobalt center along the z-axis. The X-ray structures of
[Tp′′]CoI and [PhTttert-butyl]CoX also show a rather strong
distortion in the axial direction (the average I-Co-N angle is
∼122° for [Tp′′]CoI; the average Cl-Co-S angle is ∼118°
for [PhTttert-butyl]CoCl and ∼118° for [PhTttert-butyl]CoI). Geo-
metric constraints for the facially capping [PhBP3], [Tp′′], and
[PhTttert-butyl] donor ligands enforce the large X-Co-L (L )
P, N, and S donor, respectively) bond angles in each case. As
previously mentioned, low spin 1 is somewhat distorted from
C3 symmetry by elongation of one of the phosphine donors.
A similar distortion is not observed in [Tp′′]CoI, nor in
[PhTttert-butyl]CoCl and [PhTttert-butyl]CoI. In these three cases,
the three Co-L (L ) N, S) bond lengths are very similar and
the three X-Co-L bond angles show little variation. The solid-
state structure of 1 is a reasonable consequence of its low spin
configuration, expected to give rise to a Jahn-Teller distortion.
That the related complexes [Tp′′]CoI, [PhTttert-butyl]CoCl, and
[PhTttert-butyl]CoI do not adopt a low spin configuration and show
a similar Jahn-Teller distortion likely reflects the relative
ligand-field donor strengths of the three borate-derived donor
ligands. The upper lying e-set (in C3V) is σ*. Its degree of
destabilization should directly reflect the relative donor strength
of three donor ligands. The [PhBP3] ligand appears to be the
strongest donor, destabilizing the upper orbital set to the greatest
extent, but by a relative amount that has yet to be quantified.
The ability of each ligand to accommodate a Jahn-Teller
distortion (if a low spin configuration is to occur), however, is
also a variable that needs to be considered. Tris(pyrazolyl)borate
ligands, when coordinated in an η3-fashion, are not very flexible.
Geometric JT-distortion to a stable low spin configuration might
therefore be energetically unfavorable. Thus, [Tp]-derived
ligands could provide a ligand-field strength similar to [PhBP3],
but still be coordinatively less flexible, thereby raising the energy
of the doublet state above the quartet state for [Tp]CoX
derivatives. Riordan’s [PhTttert-butyl]CoX system is another
matter. That the monomeric complexes [PhTttert-butyl]CoX are
high spin is somewhat surprising based on the electronic scheme
presented herein. It would seem that the tris(thioether)borate
ligands are flexible enough to permit the Jahn-Teller distortion.
The high-spin character of the [PhTttert-butyl]CoX derivatives is
perhaps reflective of their less pronounced axial elongation and
their attenuated ligand-field strength by comparison to the
[PhBP3]CoX systems. This would have the effect of narrowing
the gap between the a1 orbital and the upper e manifold, thereby
making the high spin configuration of (PhTttert-butyl)CoX com-
plexes more stable. It would certainly be of interest to
Figure 11. Figure shows the experimental (X-ray) (A) and calculated (DFT)
(B) molecular structures for complex 1. Representations of the HOMO-1
(C) and HOMO (D) orbitals obtained from the DFT electronic structure
calculations are also shown. The calculated (DFT) bond distances (Å) and
angles (deg) were theoretically determined to be as follows. The experi-
mentally determined value is shown in parentheses: Co-I, 2.585 (2.474);
Co-P(1), 2.282 (2.163); Co-P(2), 2.240 (2.208); Co-P(3), 2.402 (2.244);
P(1)-Co-P(2), 88.6 (90.5); P(1)-Co-P(3), 96.0 (92.9); P(2)-Co-P(3),
90.5 (92.3); P(1)-Co-I, 119.0 (119.6); P(2)-Co-I, 131.6 (127.3); P(3)-
Co-I, 121.6 (124.6).
distortion is so severe that the angle (L-Co-L) between the
three fac ligands becomes 90°, enforced in complex 1 by the
[PhBP3] ligand, the set of orbitals derived for Oh symmetry
correlate very well with the axially distorted pseudo-tetrahedral
structure type. For a low spin configuration we anticipate a
Jahn-Teller distortion from C3V to Cs symmetry. The lower
portion of Figure 10 depicts this situation and suggests that the
final ground electronic configuration (using the functions given
above) to be ({(2/3)1/2(x2 - y2) - (1/3)1/2yz})2({(2/3)1/2xy -
(1/3)1/2xz})2(z2)2({(1/3)1/2x2 - y2 + (2/3)1/2yz})1({(1/2)1/2xy +
(2/3)1/2xz})0. The EPR spectrum of 1 (g ) 2.20 and g ) 2.05)
||
is consistent with this configuration to the extent that g is
||
appreciably higher than the free electron value, suggesting that
the unpaired spin is not localized in an orbital of z2 parentage.
Rather, the unpaired spin resides in an orbital that is formally
{(1/3)1/2x2 - y2 + (2/3)1/2yz} within the present coordinate axis
scheme.
Hoping to corroborate this conclusion we performed a
theoretical electronic structure calculation on the title complex
1. The geometry of the complete complex, including the aryl
rings of the [PhBP3] ligand, was theoretically determined by
performing a geometry optimization that used the experimentally
determined structural coordinates for 1 as a starting point (see
the Experimental Section for details). The structure calculation
converged and reasonable, though not excellent, agreement
between the theoretical structure of 1 and its experimentally
determined structure was established (see Figure 11, parts a and
9
J. AM. CHEM. SOC. VOL. 124, NO. 51, 2002 15347

A R T I C L E S
Jenkins et al.
comparatively study the temperature dependence of the mag-
netic behavior of Theopold’s and Riordan’s cobalt(II) systems
by SQUID magnetometry to establish whether any of the low
spin form for these complexes is populated at very low
temperatures.
related systems that have been previously studied. Two scenarios
that transform a high spin, 4-coordinate tetrahedral complex to
a 4-coordinate species that is low spin can therefore be put
forward. The first scenario pertains to the traditional case in
which a high spin tetrahedral structure is distorted to a low-
spin square planar structure type. The second scenario, that
which is described herein, occurs when coupling of a strong
ligand-field donor strength with a pronounced axial distortion
leads to a distorted tetrahedral geometry that can accommodate
a cobalt(II) ion in a low spin ground-state configuration. The
latter, more gentle distortion exposes an orbital at the cobalt(II)
center appropriate for small molecule binding, a feature that
could be exploited if tunable crossover systems for 4-coordinate
cobalt(II) prove accessible. We are pursuing this possibility by
searching for new [PhBP3]CoX systems that display spin-cross-
over phenomena while maintaining 4-coordination in an ap-
proximate tetrahedral structure type. We are also pursuing new
tris(phosphino)borate ligands to further test our electronic model.
An interesting observation for the low spin complexes 1, 2,
and 3 is that they bind a stoichiometric equivalent of CO readily
to form the low spin, 5-coordinate carbonyl adducts [PhBP3]-
Co(CO)(X) (X ) I-, Br-, Cl-). Although a full discussion of
these and related 5-coordinate [PhBP3]Co(II) complexes will
be provided elsewhere,43 it is of interest in the present discussion
to compare the relative CO donor strength of [PhBP3]Co(CO)-
(Br) to the isostructural but cationic complex [{CH3C(CH2-
PPh2)3}Co(Br)(CO)][BPh4] previously reported by Sacconi.23d
The carbonyl stretching frequency of structurally characterized
[PhBP3]Co(CO)(Br) is 2030 cm-1. The carbonyl stretching
frequency of structurally characterized [{CH3C(CH2PPh2)3}-
Co(Br)(CO)][BPh4] is 2060 cm-1. That the cationic complex
exhibits a CO stretch 30 cm-1 higher in energy than the neutral
derivative suggests that the anionic borato ligand is an ap-
preciably stronger donor than its neutral ligand counterpart. The
increased donor strength likely accounts for the tendency of
dimeric bromide 2 to dissociate to a 15 electron, low spin
monomer in solution. Such an equilibrium has not been observed
for any of the cationic cobalt(II) dimers supported by triphos,
in which a halide or pseudohalide occupies the bridging
position.23,24 We highlight this distinction because it is important
to bear in mind with respect to the degree of “zwitterionic”
character ascribed to neutral complexes supported by (phos-
phino)borate ligands. From the perspective of a simple Lewis
structure depiction, simple resonance contributors that delocalize
the anionic borate charge to the phosphine donor ligands are
not available. This contrasts the situation for tris(pyrazolyl)-
borate ligands, where simple resonance delocalization is ex-
pected to partially distribute the negative charge to the N-donor
atoms. That [PhBP3] is such a strong field donor, however,
suggests that there is appreciable communication between the
borate anion and the bound metal center, either through space
or through the sigma framework. Our ability to identify low
spin, monomeric cobalt(II) halides that do not irreversibly
dimerize might be a direct consequence of the strong field of
the [PhBP3] anion: it stabilizes complexes that are formally
electron deficient. The high spin complexes 4 and 5 are also
monomeric, as are the high spin systems of Theopold and
Riordan.26,27 This is perhaps to be anticipated, as they do not
possess an empty d orbital to accommodate the lone pair of a
bridging halide ligand. In accord with this model, we note that
the high spin cobalt(II) derivatives 4 and 5, in addition to the
Riordan and Theopold systems, appear to be more resistant to
coordination of carbon monoxide. This fact may, however,
simply reflect other factors such as relative ligand-field donor
strengths and, for the case of 5 in particular, steric consider-
ations.
Experimental Section
All manipulations were carried out using standard Schlenk or
glovebox techniques under a dinitrogen atmosphere. Unless otherwise
noted, solvents were deoxygenated and dried by thorough sparging with
N2 gas followed by passage through an activated alumina column.
Nonhalogenated solvents were typically tested with a standard purple
solution of sodium benzophenone ketyl in tetrahydrofuran to confirm
effective oxygen and moisture removal. The reagents CoI2, CoBr2,
CoCl2, and TlPF6 were purchased from commercial vendors and used
without further purification. The preparation of [PhBP3] has been
previously reported.13 TlO-2,6-Me2Ph was prepared according to a
literature procedure.44 Elemental analyses were performed by Desert
Analytics, Tucson, AZ. A Varian Mercury-300 NMR spectrometer or
1
a Varian Inova-500 NMR spectrometer was used to record H NMR
1
spectra at ambient temperature unless otherwise stated. H chemical
shifts were referenced to residual solvent. Deuterated toluene and
benzene were purchased from Cambridge Isotope Laboratories, Inc.
and were degassed and dried over activated 3 Å molecular sieves prior
to use. UV-vis measurements were taken on a Hewlett-Packard 8452A
diode array spectrometer using a quartz crystal cell with a Teflon cap.
Near-IR measurements were taken on a Cary 14 spectrophotometer
using a quartz crystal cell with a Teflon cap. IR measurements were
obtained using a Bio Rad Excalibur FTS 3000 with a KBr solution
cell. X-ray diffraction studies were carried out in the Beckman Institute
Crystallographic Facility on a Bruker Smart 1000 CCD diffractometer.
Magnetic Measurements. Measurements were recorded using a
Quantum Designs SQUID magnetometer running MPMSR2 software
(Magnetic Property Measurement System Revision 2). Data were
recorded at 5000 G. Samples were suspended in the magnetometer in
plastic straws sealed under nitrogen with Lilly No. 4 gel caps. Each
measurement was performed on samples that had been recently
subjected to combustion analysis to verify their purity. Loaded samples
were centered within the magnetometer using the DC centering scan
at 35 K and 5000 gauss. Data were acquired at 2-10 K (one data point
every 2 K), 10-60 K (one data point every 5 K), 60-310 K (one data
point every 10 K).
øM
øm )
(1)
µeff
)
7.997ø T
(2)
x
m
IV. Conclusion
mG
In summary, we have synthesized and studied a series of
divalent cobalt complexes supported by the [PhBP3] anion.
Structural, magnetic, and spectroscopic data for each of these
systems is suggestive of low spin behavior for their monomeric,
distorted tetrahedral structure types. The low spin nature of the
monomeric form of these halides appears to be unique from
The magnetic susceptibility was adjusted for diamagnetic contributions
using the constitutive corrections of Pascal’s constants and a fixed
temperature independent paramagnetism (TIP) crudely set to 2 × 10-4
cm3 mol-1 45
The molar magnetic susceptibility (øm) was calculated
.
by converting the calculated magnetic susceptibility (ø) obtained from
the magnetometer to a molar susceptibility (using the multiplication
9
15348 J. AM. CHEM. SOC. VOL. 124, NO. 51, 2002

Low Spin Cobalt(II) System
A R T I C L E S
factor {(molecular weight)/[(sample weight)(Field Strength)]}). Curie-
Weiss behavior was verified by a plot of øm versus T (shown in
7.7 (s), 7.5 (s), 4.3 (bs), 2.2 (s), -8.5 (s). IR (cm-1): 1433, 1091, 739.
UV-vis (C6H6) λmax, nm (ꢀ): 638 (1112), 738 (627). SQUID (solid -
average 30-220K): µeff ) 2.65 ( 0.03 µB Θ ) -1.43 K. Evans
Method (C6D6): 2.76 µB. Anal. Calcd for C45H41BCoIP3: C, 62.03; H,
4.74. Found: C, 61.76; H, 4.75.
-1
Figures 2b and 3b). Data were analyzed using eqs 1 and 2. Average
magnetic moments were taken from the average of magnetic moments
from the ranges indicated in the Experimental Section for each complex.
The Weiss constant (Θ) was taken as the x-intercept of the plot of
Synthesis of {[PhBP3]Co(µ-Br)}2, 2. A benzene solution (50 mL)
of the thallium reagent, [PhBP3]Tl (0.507 g, 0.570 mmol), was added
to a stirring suspension of CoBr2 (0.250 g, 1.14 mmol) in benzene (20
mL). After stirring at ambient temperature for 24 h, the resulting green
solution was filtered through diatomaceous earth. The filtrate was
pumped to dryness to leave a purple powder that was redissolved in
benzene (20 mL), filtered over diatomaceous earth, and again thoroughly
dried in vacuo. A UV-vis spectrum of this powder suggested a mixture
of products. The powder was redissolved in benzene (5 mL) and filtered
over a fine sintered-glass frit. A microcrystalline purple powder
precipitated from the solution; this powder (0.132 g, 28.1%) was dried
and analyzed: 1H NMR (C6D6, 300 MHz): δ 26.7 (bs), 11.0 (s), 7.6
(s), 7.4 (s), 4.2 (bs), 2.2 (s). UV-vis (C6H6) λmax, nm (ꢀ): 612 (1773),
718 (958). SQUID (solid, average 10-100 K): µeff ) 1.77 ( 0.08 µB,
Θ ) -1.48 ( 0.49 K; (solid, average 110-310 K) 2.11 µB. Evans
Method (C6D6, calculated for a monomer): 2.6 µB. Anal. Calcd for
C90H82B2Br2Co2P6: C, 65.56; H, 5.01. Found: C, 65.23; H, 4.87. The
sample used for the EPR and SQUID measurements, which was
recrystallized two more times, analyzed as follows: C, 65.70; H, 5.04.
Synthesis of {[PhBP3]Co(µ-Cl)}2, 3. A benzene solution (50 mL)
of the thallium reagent, [PhBP3]Tl (0.857 g, 0.963 mmol), was added
to a stirring suspension of CoCl2 (0.250 g, 1.93 mmol) in benzene (20
mL). After stirring at ambient temperature for 24 h, the resulting green
solution was filtered through diatomaceous earth. The filtrate was dried
in vacuo to a purple powder that was then redissolved in benzene (20
mL), filtered over diatomaceous earth, and again thoroughly dried. The
resulting purple powder was redissolved in benzene (5 mL) and filtered
over a fine sintered-glass frit; this caused a microcrystalline purple
powder to precipitate from solution. The dried powder (0.195 g, 25.9%)
was analyzed: 1H NMR (C6D6, 300 MHz): δ 31.9 (bs), 15.0 (s), 11.2
(s), 9.1 (s), 7.7 (s), 7.4 (s), 3.9 (bs), 2.2 (s). UV-vis (C6H6) λmax, nm
(ꢀ): 594 (1507), 712 (724). SQUID (solid - average 10-100 K): µeff
) 2.47 ( 0.14 µB, Θ ) -3.13 ( 0.67 K; (solid - average 110-310
K) 2.79 µB. Evans Method (C6D6, calculated for a monomer): 2.8 µB.
Anal. Calcd for C90H82B2Cl2Co2P6: C, 69.30; H, 5.30. Found: C, 68.14;
H, 5.18. The sample used for the EPR and SQUID measurements, which
was recrystallized two more times, analyzed as follows: C, 69.56; H, 5.36.
Alternative Preparation for 2 and 3. Thallium hexafluorophosphate
(0.090 g, 0.26 mmol) was added to a stirring THF (20 mL) solution of
1 (0.201 g, 0.231 mmol). A yellow powder precipitated within 5 min
and was subsequently removed by filtration through diatomaceous earth.
NaCl (0.26 g, 4.4 mmol) was added to the filtrate, and the reaction
solution was allowed to stir for 24 h. The above process was then
repeated with additional TlPF6 and NaCl. After another 24 h, the salts
were removed by filtration; a UV-vis spectrum of the product solution
indicated that 1 had been fully consumed, and the spectrum matched
that of 3. The solution was concentrated in vacuo to 50% of its original
volume and allowed to stand for 24 h. White precipitate (NaCl) was
removed by filtration, and the green solution was concentrated to
dryness, redissolved in benzene, and once again allowed to stand for
12 h. Repeated filtration through diatomaceous earth ensured complete
removal of salts. The final product was obtained by drying thoroughly
in vacuo to afford purple 3 as a fine powder (0.159 g, 88.6%). An
analogous protocol was effective for converting 1 to 2, in this case
using KBr instead of NaCl. Both UV-vis and 1H NMR spectroscopies
are useful in monitoring the metathesis procedure to ensure complete
conversion.
-1
øm versus T. Error bars were established at 95% confidence using
regression analysis or taking two standard deviations from the mean.
Solution magnetic moments were measured by the method of Evans
and were adjusted for diamagnetic contributions using the constitutive
corrections of Pascal’s constants.29
Averaged g-factors can be extracted from the susceptibility data,
assuming zero orbital contributions, using the following equation:45
Ng2â2
3kT
øm )
(S(S + 1))
EPR Measurements. X-band EPR spectra were obtained on a
Bruker EMX spectrometer equipped with a rectangular cavity working
in the TE102 mode. Variable temperature measurements were conducted
with an Oxford continuous-flow helium cryostat (temperature range
3.6-300 K). Accurate frequency values were provided by a frequency
counter built in the microwave bridge. Solution spectra were acquired
in toluene for all of the complexes. Sample preparation was performed
under a nitrogen atmosphere, particularly important for handling the
low spin complexes 1, 2, and 3 (note: adventitious oxygen is
problematic because it gives rise to a high spin monomer concentration,
as in the conversion of 1 to 4).
EPR simulations for the monomers were performed with the program
WINEPR SimFonia (Version 1.25, Bruker Analytische Messtechnik
GmbH); this software is based on second-order perturbation solution
of the spin Hamiltonian: H ) H‚g‚S + Σ S‚A‚I. For the dimers, the
absolute values of the spin-Hamiltonian parameters were extracted from
simulations of the EPR spectra. Calculations were carried out on a PC
using FORTRAN code based on Gladney’s general EPR fitting
program,46 and adapted for simulations of EPR spectra of randomly
oriented spin triplets. Transition fields and corresponding average
transition moments47 were computed in the magnetic field domain48
by matrix diagonalization of the S ) 1 spin Hamiltonian.
3
2
Hˆ ) â(gzHzSz + gxHxSx + gyHySy) + D S2z - + E(Sx2 - Sy2)
(
)
D and E are defined in terms of the components of the diagonal fine-
structure tensor as follows:
2
3
1
3
1
3
Dzz ) D; Dxx ) - D + E; Dyy ) - D - E
DFT Calculations. A geometry optimization was performed for
complex 1 using Jaguar (Version 4.1) starting from coordinates based
on the solid-state structure that had been determined by X-ray
diffraction. No symmetry constraints were imposed and the calculation
was performed on a doublet electronic state. The method used was
B3LYP with LACVP** as the basis set (modified to LACVP**++
for the boron atom). Optimization was considered converged when
energy changes in successive iterations fell below 0.3 kcal/mol. It should
also be noted that convergence was also achieved for a system when
the geometry was optimized assuming a quartet ground state.
Synthesis of [PhBP3]CoI, 1. A benzene solution (50 mL) of the
thallium reagent, [PhBP3]Tl, (0.356 g, 0.400 mmol) was added to a
stirring suspension of CoI2 (0.250 g, 0.799 mmol) in benzene (20 mL).
After stirring at ambient temperature for 24 h, the resulting green
solution was filtered through diatomaceous earth, concentrated in vacuo
(50% original volume), and filtered through diatomaceous earth once
again. Vapor diffusion of petroleum ether into the resulting green filtrate
afforded the pure crystalline product (0.317 g, 91.1%) which was
analyzed: 1H NMR (C6D6, 300 MHz): δ 22.2 (bs), 15.8 (s), 10.8 (s),
Synthesis of [PhB(CH2P(O)Ph2)2(CH2PPh2)]CoI, 4. A 100 mL
Schlenk flask with a Teflon stir bar was charged with a benzene (30
mL) solution of 1 (0.278 g, 0.319 mmol) and stirred at room temper-
ature. Oxygen (7.8 mL at 1 atm, 0.32 mmol) was added to the reaction
vessel via syringe through a rubber septum. After 4 h, the solution,
9
J. AM. CHEM. SOC. VOL. 124, NO. 51, 2002 15349

A R T I C L E S
Jenkins et al.
which had turned bright blue, was dried in vacuo to a fine blue powder.
Washing this powder with petroleum ether (3 × 10 mL) and concen-
trating it to dryness afforded the final product (0.225 g, 79.1% yield)
The largest peak and hole in the difference map were 0.821 and -0.495
e.Å-3, respectively. Maximum and minimum transmission were equal
to 0.957 and 0.877, respectively. Crystal data for C90H82B2Cl2Co2P6‚
(2C6H6), monoclinic space group P21/c (#14), a ) 26.340(3) Å, b )
13.2151(14) Å, c ) 28.015(3) Å, b ) 117.909(2)°, V ) 8617.6(16)
Å3, Z ) 4, Dcalcd ) 1.323 g/cm3, Abs. Coefficient ) 0.607 mm-1, Mo-
KR λ ) 0.710 73 Å, T ) 98 K, Bruker SMART 1000 CCD, crystal
size 0.22 × 0.19 × 0.07 mm3, θmax ) 28.13°, R1 ) 0.0468, wR2 )
0.0660 for I > 2s(I), R1 ) 0.0995, wR2 ) 0.0724, GF (1/σ2 weighting)
) 1.028, number of reflections collected 106 341 (-34 ) h ) 33,
-16 ) k ) 17, -37 ) l ) 36), number of independent reflections
20 585, number of parameters 1027.
X-ray Crystal Structure Analysis of 4. X-ray quality crystals were
obtained by vapor diffusion of petroleum ether into a concentrated
solution of 4 in benzene. An electric blue crystalline shard was mounted
on a glass fiber with Paratone N oil. The structure was solved by direct
methods in conjunction with standard difference Fourier techniques.
The largest peak and hole in the difference map were 1.060 and -1.086
e.Å-3, respectively. Maximum and minimum transmission were equal
to 0.906 and 0.810, respectively. Crystal data for C57H53BCoIO2P3,
monoclinic space group P21/n (#14), a ) 12.916(3) Å, b ) 18.582(4)
Å, c ) 21.674(5) Å, â ) 106.262(3)°, V ) 4993.7(18) Å3, Z ) 4,
Dcalcd ) 1.409 g/cm3, Abs. Coefficient ) 1.100 mm-1, Mo-KR λ )
0.71073 Å, T ) 98 K, Bruker SMART 1000 CCD, crystal size 0.22 ×
0.09 × 0.16 mm3, θmax ) 23.32°, R1 ) 0.0361, wR2 ) 0.0602 for I >
2s(I), R1 ) 0.0547, wR2 ) 0.0637, GF (1/σ2 weighting) ) 1.755,
number of reflections collected 54319 (-14 ) h ) 14, -20 ) k ) 20,
-24 ) l ) 24), number of independent reflections 7213, number of
parameters 586.
1
that was analyzed. H NMR (C6D6, 300 MHz): δ 16.2 (s), 13.8 (s),
10.1 (s), 9.47 (s), 9.28 (s), 7.90 (s), 7.09 (s), -2.1 (s). UV-vis (C6H6)
λ
max, nm (ꢀ): 590 (450), 642 (461), 682 (520). IR (cm-1): 1435, 1126,
1095, 1071, 752. SQUID (solid, average 30-310 K): µeff ) 4.33 (
0.09 µB, Θ ) -2.74 K. Evans Method (C6D6): 4.4 µB. Anal. Calcd
for C45H41BCoIP3O2: C, 59.83; H, 4.57. Found: C, 59.62; H, 4.80.
Synthesis of [PhBP3]Co(O-2,6-dimethylphenyl), 5. A THF solution
(5 mL) of the thallium reagent Tl-O-2,6-dimethylphenyl (0.0433 g,
0.133 mmol) was added to a stirring solution (10 mL) of 1 (0.116 g,
0.133 mmol). A yellow precipitate (TlI) formed within 15 min and
was removed by filtration through diatomaceous earth. The remaining
brown solution was stirred overnight. The reaction volatiles were
removed in vacuo and the resulting brown powder was redissolved in
benzene. The benzene was lyophilized to thoroughly remove any
remaining THF. The resulting fine powder was filtered once more
through diatomaceous earth as a benzene (3 mL) solution. Vapor
diffusion of petroleum ether into this benzene solution afforded the
red-brown, crystalline product (0.067 g, 58.7% yield) that was analyzed.
1H NMR (C6D6, 300 MHz): δ 67.0 (s), 52.4 (s), 17.9 (s), 12.4 (s),
11.6 (s), 11. 0 (s), 9.0 (s), 8.45 (s), 7.80 (s), 7.45 (s), 7.06 (s), -3.4
(bs), -6.5 (s), -62.2 (s). UV-vis (C6H6) λmax, nm: 534 (1740), 752
(624). SQUID (solid, average 30-310 K): µeff ) 4.36 ( 0.18 µB, Θ
) 7.48 K. Evans Method (C6D6): 4.4 µB. Anal. Calcd for C53H50-
BCoP3O: C, 73.54; H, 5.82. Found: C, 73.32; H, 5.95.
X-ray Crystal Structure Analysis of 1. X-ray quality crystals were
obtained by vapor diffusion of petroleum ether into a concentrated
solution of 1 in benzene. A dark green crystalline blade was mounted
on a glass fiber with Paratone N oil. The structure was solved by direct
methods in conjunction with standard difference Fourier techniques.
The largest peak and hole in the difference map were 0.775 and -0.614
e.Å-3, respectively. Maximum and minimum transmission were equal
to 0.895 and 0.804, respectively. Crystal data for C45H41BCoIP3‚
1/2(C6H6), monoclinic space group P21/c (#14), a ) 22.5443(17) Å,
b ) 12.7044(9) Å, c ) 29.526(2) Å, â ) 90.230(2)°, V ) 8456.6(11)
Å3, Z ) 8, Dcalcd ) 1.430 g/cm3, Abs. Coefficient ) 1.274 mm-1, Mo-
KR λ ) 0.710 73 Å, T ) 98 K, Bruker SMART 1000 CCD, crystal
size 0.18 × 0.13 × 0.09 mm3, θmax ) 25.55°, R1 ) 0.0476, wR2 )
0.0786 for I > 2s(I), R1 ) 0.0899, wR2 ) 0.0843, GF (1/σ2 weighting)
) 1.137, number of reflections collected 50716 (-24 ) h ) 23, -13
) k ) 13, -31 ) l ) 31), number of independent reflections 11 064,
number of parameters 926.
X-ray Crystal Structure Analysis of 2. X-ray quality crystals were
obtained by vapor diffusion of petroleum ether into a concentrated
solution of 2 in benzene. A purple crystalline needle was mounted on
a glass fiber with Paratone N oil. The structure was solved by direct
methods in conjunction with standard difference Fourier techniques.
The largest peak and hole in the difference map were 2.231 and -1.641
e.Å-3, respectively. Maximum and minimum transmission were equal
to 0.915 and 0.713, respectively. Crystal data for C90H82B2Br2Co2P6,
orthorhombic space group Pbcn (#50), a ) 22.214(3) Å, b ) 19.286(3)
Å, c ) 18.008(3) Å, V ) 7715.0(19) Å3, Z ) 4, Dcalcd ) 1.419 g/cm3,
Abs. Coefficient ) 1.637 mm-1, Mo-KR λ ) 0.71073 Å, T ) 98 K,
Bruker SMART 1000 CCD, crystal size 0.22 × 0.11 × 0.05 mm3,
θmax ) 27.78°, R1 ) 0.0695, wR2 ) 0.0965 for I > 2s(I), R1 ) 0.1432,
wR2 ) 0.1012, GF (1/σ2 weighting) ) 1.534, number of reflections
collected 9263 (0 ) h ) 29, 0 ) k ) 26, 0 ) l ) 24), number of
independent reflections 9263, number of parameters 461.
X-ray Crystal Structure Analysis of 5. X-ray quality crystals were
obtained by vapor diffusion of petroleum ether into a concentrated
solution of 5 in benzene. A purple crystal was mounted on a glass
fiber with Paratone N oil. The structure was solved by direct methods
in conjunction with standard difference Fourier techniques. The largest
peak and hole in the difference map were 1.133 and -0.475 e.Å-3
,
respectively. Maximum and minimum transmission were equal to 0.913
and 0.902, respectively. Crystal data for C53H50BCoOP3, monoclinic
space group Cc (#9), a ) 17.245(2) Å, b ) 16.484(1) Å, c ) 16.494(1)
Å, â ) 109.181(1)°, V ) 4428.3(6) Å3, Z ) 4, Dcalcd ) 1.298 g/cm3,
Abs. Coefficient ) 0.535 mm-1, Mo-KR λ ) 0.710 73 Å, T ) 96 K,
Bruker SMART 1000 CCD, crystal size 0.170 × 0.185 × 0.192 mm3,
θmax ) 28.61°, R1 ) 0.0378, wR2 ) 0.0612 for I > 2σ(I), R1 ) 0.0454,
wR2 ) 0.0624, GF (1/σ2 weighting) ) 1.661, number of reflections
collected 45 708 (-23 ) h ) 22, -21 ) k ) 21, -21 ) l ) 22),
number of independent reflections 10 456, number of parameters 534.
Acknowledgment. We are grateful to the Dreyfus Foundation,
the ACS Petroleum Research Fund, and the National Science
Foundation (Grant No. CHE-0132216) for financial support of
this work. We thank the Beckman Institute (Caltech) for use of
the SQUID magnetometer and the crystallographic facility.
D.M.J. is grateful for a pre-doctoral fellowship from the National
Science Foundation. M.J.A. is grateful for a NDSEG Fellowship.
T.A.B. is grateful for a pre-doctoral fellowship from the DOD.
J. C. Thomas is acknowledged for assistance with DFT. Finally,
we acknowledge our inorganic colleagues at Caltech for many
stimulating discussions, and the reviewers of this paper for
helpful commentary.
Supporting Information Available: 1H NMR spectra for all
complexes; EPR spectrum of 5; crystallographic data. This
material is available free of charge via the Internet at
http://pubs.acs.org.
X-ray Crystal Structure Analysis of 3. X-ray quality crystals were
obtained by vapor diffusion of petroleum ether into a concentrated
solution of 3 in benzene. A purple crystalline block was mounted on
a glass fiber with Paratone N oil. The structure was solved by direct
methods in conjunction with standard difference Fourier techniques.
JA026433E
9
15350 J. AM. CHEM. SOC. VOL. 124, NO. 51, 2002